Journal of Industrial and Engineering Chemistry 92 (2020) 109–119

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry

journal homepage: www.elsevier.com/locate/jiec

Kinetics for the biodiesel production from lauric acid over Keggin
heteropolyacid loaded in silica framework
Luis A. Gallego-Villadaa , Edwin A. Alarcóna,* , Valeria Palermob , Patricia G. Vázquezb ,
Gustavo P. Romanellib,c
a
Chemical Engineering Department, Environmental Catalysis Research Group, Universidad de Antioquia, Calle 70 No. 52-21, Medellín 050010, Colombia
b
Centro de Investigación y Desarrollo en Ciencias Aplicadas “Dr. Jorge J. Ronco” CINDECA, (CONICET-CIC-UNLP), Facultad de Ciencias Exactas, Universidad
Nacional de La Plata, Calle 47 No. 257, B1900AJK La Plata, Argentina
c
Cátedra de Química Orgánica, Facultad de Ciencias Agrarias y Forestales, Universidad Nacional de La Plata, Calles 60 y 119 s/n, B1904AAN La Plata,
Argentina

A R T I C L E I N F O A B S T R A C T

Article history: Kinetic models were developed to describe the esterification reaction of lauric acid and ethanol over
Received 26 June 2020 heterogeneous catalysts, which is a reaction of special interest in the biodiesel production. Vanadium
Received in revised form 25 August 2020 Keggin heteropolyacid was included on a silica framework by sol–gel procedure, using different loadings.
Accepted 28 August 2020
The synthesized materials were characterized by FT-IR, XRD, SEM, nitrogen adsorption/desorption
Available online 6 September 2020
isotherms, and potentiometric titration, and tested as solid catalysts in the esterification of lauric acid.
Best performance was achieved with SiO2@20PMoV, which was used in the esterification of others fatty
Keywords:
acid and alcohols. The reuse was successfully tested in five consecutive runs. Kinetic data using
Kinetic modeling
Esterification
SiO2@20PMoV were obtained at different temperatures (48–78
C), fatty acid:alcohol molar ratios
Lauric acid (4 mmol of lauric acid and 2.5, 5, 10 and 15 mL of ethanol), and catalyst amounts (12.5, 25 and 50 mg). The
Heterogeneous catalysis best heterogeneous models were LH3 (surface reaction as rate-limiting step) and ER3 (desorption of ethyl
Biodiesel laurate as rate-limiting step when the adsorbed reactant is lauric acid). The activation energies were
Vanadium Keggin heteropolyacid 56.1 kJ mol1 and 64.8 kJ mol1, and the reaction rate constants at 78
C were 0.2791 mol g1 h1 and
0.0768 mol g1 h1 for LH3 and ER3, respectively.
© 2020 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights
reserved.

Introduction
Biomass is a potential alternative to non-renewable fossil fuels
for the future, due to aggravation of the energy crisis [1] and to that
its use as feedstock reduces the CO2 content in the atmosphere [2].
Biodiesel is a renewable fuel consist of long-chain esters formed
from fatty acids and methanol or ethanol that can be obtained
either by transesterification reaction of vegetable oils and fats or
esterification reaction of fatty acids present in animal fats (e.g. lard
or tallow) [3]; contrary to fossil-based diesel, biodiesel also offers
biodegradability and non-toxicity [4]. The key factor of the
globalization of biodiesel is its sustainability with available diesel
engines without the need of mechanical modifications [5].
The esterification of carboxylic acids with alcohols or phenols is
a useful preparation method of long-chain organic esters, which
have many industrial applications, especially as biofuels [6].
* Corresponding author.
E-mail address: edwin.alarcon@udea.edu.co (E.A. Alarcón).
Coconut and palm kernel oils are the major source of medium-
chain fatty acids feedstocks, specifically with high content of lauric
acid (dodecanoic acid, C12:0) [7,8]. Traditionally, esterification of
fatty acid with alcohol is carried out over homogeneous mineral
acid catalyst. Catalysts commonly used are: sulfuric acid, hydro-
chloric acid and organic acids like p-toluenesulfonic acid [9].
However, there are other types of the suitable catalysts for
biodiesel synthesis as the heterogeneous or enzymatic catalysts,
which offer easy separation of products and less wastewater,
avoiding environmental issues and corrosion [10–12], and allowing
continuous operating of reactors and catalyst reuse [13,14].
Different compounds have been tested as heterogeneous
catalysts for biodiesel production: ionic exchange resins, sup-
ported-metal compounds, zeolites, and polymers are some
representative examples [15–17]. In the last decades, heteropo-
lyacids (HPAs) have emerged as sustainable catalysts in Organic
Chemistry due to these have been recognized as economically and
environmentally benign acid catalysts due to their acidity and
redox propertier for various reactions [18]. We have developed
new materials based in them, for the use in the synthesis of

https://doi.org/10.1016/j.jiec.2020.08.030
1226-086X/© 2020 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
110 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119
heterocycles, eco-compatible oxidations with hydrogen peroxide,
and recently in the valorization of biomass derivatives [19–22].
The major disadvantages of bulk HPAs derivatives are the low
specific area (1–10 m2 g1), low thermal stability and the high
solubility in polar media [23]. For this reason and in order to
overcome these problems, different supports were employed to
immobilize HPAs, for example activated carbon, zeolites, silica,
zirconium, titanium, MCM-41, and polymers among others
[24,25]. Different HPAs in heterogeneous form have been tested
in biodiesel production such as cesium-doped heteropoly
tungstate (HPW), HPW/ZrO2, HPW/g-Al2O3 and HPW/SiO2 for
the simultaneous esterification and transesterification of 10%
oleic acid-soybean oil mixture [26]; core–shell nanostructured
heteropoly acid-functionalized zeolitic imidazolate frameworks-
8 (ZIF-8) for rapeseed oil transesterification [18]; HPW/kaolinite,
HPW/bentonite and HPW/montmorillonite for acetic acid esteri-
fication [27]; HPW/SiO2 and cesium-doped HPW for rapeseed oil
transesterification [28]; H3PMo12O40/bentonite [29] for esterifi-
cation of a waste from palm oil; H3PW12O40/KIT-6 for neem oil
transesterification [30]; 12-tungstophosphoric HPA/ZrO2 [31],
and H3PW12O40 and H3PMo12O40 supported on activated carbon
fibers [32] for palmitic acid esterification; Ni0.5H3SiW/SiO2 [33],
NiHSiW/UiO-66 [34] and Sn1.5PW12O40/Cu-BTC [35] for oleic acid
esterification.
Pseudo-homogeneous kinetic models have been used to describe
the biodiesel production from the lauric acid esterification with
ethanol over several catalytic systems such as ZnL2 [36], [(n–bu–
SO3H)MIM][HSO4] [37], deep eutectic solvents based on cetyl
trimethyl ammonium bromide (CTAB) [38], and over 1-methyl-2-
pyrrolidonium hydrogen sulfate ([Hnmp]HSO4) [39] and silico-
tungstic acid encapsulated UiO-66 [40] using methanol. However,
Langmuir–Hinshelwood–Hougen–Watson (LHHW) and Eley–
Rideal (ER) heterogeneous kinetic models have been succesfully
used to describe esterification kinetics of several heterogeneous
catalytic systems for biodiesel production over niobium oxide [41],
Amberlyst 46 [42], ion exchange resin [43], Amberlyst 15 [44] and
12-tungstophosphoric acid/SBA-15 [45]. Particularly for lauric acid
esterification with ethanol, a heterogeneous kinetic study by Eley–
Rideal mechanism over acid activated montmorillonite has been
reported [46], which is a promising catalyst because of their low
cost, thermal stability, reusability and mainly because their
desirable catalytic properties such as high selectivity, high surface
area, high pores dimension and presence of acid sites; furthermore,
this catalyst allows to achieve a high equilibrium conversion around
90%. The experimental data of the heterogeneous reaction was
successfully fitted to Eley–Rideal model with surface reaction
between adsorbed ethanol and lauric acid as the rate-limiting step.
A drawback of this study is the energy intensive in reaction due to
the high temperature (180
C); furthermore, the kinetic modeling
using the LHHW mechanism-based models was not reported, which
are typical kinetic models in biodiesel production over different
heterogeneous catalytic as previously reported.
To date, there have been no reports on carrying out kinetic
studies about esterification reaction for biodiesel production over
vanadium Keggin heteropolyacids. Kinetic studies are receiving
much importance since they provide the most powerful method of
investigating the detailed reaction mechanisms, give information
about the maximum product yield and specifically in the chemical
industry, these kinetic models are the basis on which the modeling
and the design of catalytic reactors is carried out as well as the
Fig. 1. Scheme of lauric acid esterification with ethanol.
transition of production from molecular scale to macro scale [47],
as is the case of biodiesel production, which has been one of the
challenges worldwide in recent years. Thus, in this contribution,
we successfully synthesized vanadium Keggin HPA included on a
silica framework with different loads, which were tested as solid
catalysts in the lauric acid esterification. The most important aim
of this research is to develop a kinetic model to describe the
esterification reaction for biodiesel production from lauric acid and
ethanol on heterogeneous catalysts (Fig. 1).
Experimental
General
All reagents and solvents were of commercial analytical grade
and used without further treatment, supplied by Sigma-Aldrich
and Fluka.
Synthesis of catalysts: PMoV included on silica
H4PMo11VO40 (PMoV) was synthesized by a known hydrother-
mal procedure method described in the literature [48] (Supple-
mentary information).
This HPA was immobilized on silica containing four different
loadings following a sol–gel procedure described in the literature
with minor modifications [49]. A mixture of ethanol (2.95 mL),
distilled water (9 mL), and a variable amount of PMoV (126, 252,
378, and 504 mg) was added to tetraethyl orthosilicate (11.2 mL),
and stirred at 80
C, for 3 h, in a dry box in the absence of moisture.
The hydrogel obtained was dehydrated at 80
C, for 2 h, in vacuum
stove. The dried gel obtained was washed four times with ethanol
at 78
C (3  10 mL) and dried at 80
C overnight. The prepared
catalysts were named as SiO2@5PMoV, SiO2@10PMoV, SiO2@15P-
MoV, and SiO2@20PMoV, where the numbers 5, 10, 15, and 20
indicate the percentage of PMoV load.
Also silica (SiO2) was prepared following the same procedure,
but using acetic acid instead of PMoV.
Catalyst characterization
Fourier transform infrared spectra (FT-IR) of catalysts were
acquired in the 400–4000 cm1 range using pellets with KBr in a
Thermo Bruker IFS 66 FT-IR equipment.
X-ray diffraction (XRD) patterns were collected in Philips PW-
1730 device, with Cu Kα radiation (l = 1.5406 Å), 20 mA, 40 kV, and
Ni filter. Scanning angle from 5
to 60
and scanning rate of 2
per
minute were used.
Scanning electron microscopy (SEM) was recorded in a Philips
505 scanning electron microscope with an accelerating voltage of
25 eV. The solid samples were previously metallized with Au.
The specific surface area calculed by Brunauer–Emmett–Teller
method, pore volume, and the mean pore diameter of the solids
were determined by nitrogen adsorption/desorption isotherms at
196
C in a Micromeritics ASAP 2020 equipment. The samples
were previously degassed at 100
C for 12 h.
Potentiometric titration was used to determine the acidic
properties of the catalysts, employing a 794 Basic Titrino Metrohm
equipment and a double junction electrode. The solids were
suspended in acetonitrile and titrated with n-butylamine in
acetonitrile (0.025 N).
 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119 111

Catalytic esterification reaction homogeneous model. The rate law (Eq. (2)) of lauric acid can be
written as:
The selected reaction to evaluated the activity of the
dC A W

synthezised catalysts, was the esterification reaction of lauric acid ¼ rA ¼ k1 C A a C B b þ k2 C C g C D d ð2Þ
dt V rxn
with ethanol. In a test tube connected to a condenser, certain
amount of lauric acid was added in a suspension of solid catalyst (of where Ci is the molar concentration of specie i, t is reaction time, W
specific weight) in ethanol and the mixture was magnetically is the weight of catalyst, Vrxn is the reaction volume, rA is the
stirred for 24 h. formation reaction rate of lauric acid, k1 and k2 are the direct and
Three variables were analized: reaction temperature (48, 58, 68, reverse reaction constants, respectively. α, β, g and d are the
and 78
C); acid:alcohol molar ratio (4 mmol of lauric acid and 2.5, reaction orders respect to lauric acid, ethanol, ethyl laurate and
5, 10 and 15 mL of ethanol), and catalyst amount (12.5, 25 and water, respectively.
50 mg). Samples were taken at regular intervals and the catalyst Molar concentrations are related to the conversion (X) of lauric
was removed by filtration. All experiments were repeated twice. acid by Eq. (3), where Ci0 is the initial concentration of specie i, ui is
The progress of the reactions were quantified using calibration defined by Eq. (4), and gi is the stoichiometric coefficient of the
curve by GC/FID in a Shimadzu chromatograph model 2014, using a specie i in esterification reaction and it is positive for products and
FID detector and a capillary column (SPB-1, length 30 m, I.D. negative for reactants.
32 mm, and film thickness 1.00 mm). The carrier gas was high-  
purity nitrogen (≧99.999%), and the optimum conditions were as C i ¼ C A0 ui þ g i X ð3Þ
follows: splitless injection mode was used (1 mL), the injector
temperature was set at 320
C, the initial column temperature was
150
C (held for 2 min), then ramped at 20
C/min to 200
C. C i0
To demonstrate the reusability of the prepared catalyst, it was ui ¼ ð4Þ
C A0
separated by filtration after completion of the esterification
reaction, washed twice with ethanol (1 mL), dried in a vacuum The Arrhenius’ equation was used to calculated the kinetic
oven at 20
C for 24 h, and reused in a five successive runs under the constants considering the direct (k1) and reversible (k2) reaction
same conditions. equations (Eq. (5)), as follows:
 
After the optimization of reaction conditions, the esterification Ea
was extended to others fatty acids and alcohols. The products were ki ¼ Ai exp  i ð5Þ
RT
identified by GC–MS Mass Spectrometry using a HP 5971 mass
detector coupled to a HP gas chromatograph (Supplementary where Ai is the pre-exponential factor of reaction i, Eai is the
information). activation energy of reaction i, R is the ideal gas constant and T is
the temperature.
Kinetic modeling Finally, Eq. (6) is obtained by replacing Eqs. (3)–(5) in Eq. (2).
  
dX W Ea1
C A0 aþb1 ð1  X Þa ðu B  X Þ
b
For the kinetic analysis of the esterification reaction (Eq. (1)) of ¼ A1 exp 
dt V rxn RT
lauric acid with ethanol, kinetic data were carried out at different  
temperatures (48–78
C), ethanol amounts (2.5–15 mL), catalyst Ea2
A2 exp  C A0 g þd1 X g þd Þ ð6Þ
amounts (12.5–50 mg) and ten reaction times varying from 0.5 RT
until 24 h. These data were used to test several proposed kinetic The objective function (OF) presented in Eq. (7) was minimized
models, which are described below. The experimental conditions for finding the optimal parameters A1, A2, Ea1, Ea2, α, β, g, d; it was
studied in this work are presented in Table 1. done using the fmincon subroutine from Matlab1 .
A þ B$C þ D ð1Þ Nt
2
X
Where A is the fatty acid (lauric acid, C12H24O2), B is the alcohol OF ¼ X i Exp  X i Mod ð7Þ
i
(ethanol, C2H5OH), C is the ethyl ester (ethyl laurate, C14H28O2) and
D is water (H2O). X Exp
i
and X Mod
i are the experimental and model values of lauric acid
conversion and Nt is related to experimental set. It is worth to
Pseudo-homogeneous reversible model mention that the kinetic parameters were obtained from a global
A homogeneous-like model [6,36,41] was tested using the estimation by regressing of experimental kinetic data for all
experimental data, for the purpose of evaluating a possible conditions presented in Table 1.
simplification of the heterogeneous kinetics adapted to

Heterogeneous models
Several heterogeneous models were proposed to evaluate them
Table 1 in the kinetics of the esterification reaction of lauric acid with
Experimental conditions used for kinetic measurements of lauric acid esterification
ethanol. A total of 17 models were tested, which 5 of them
with ethanol, using SiO2@20PMoV as catalyst.
correspond to Langmuir–Hinshelwood–Hougen–Watson (LHHW)
Run T (
C) Lauric acid (mmol) Ethanol (mL) Catalyst (mg) kinetic models (Table 2) and 12 to Eley–Rideal (ER) kinetic models
1 58 4 10 12.5 (Table 3). All of them assume that diffusion steps were not rate-
2 58 4 10 25 limiting and that reactants do not dissociate [41]; in addition, they
3 58 4 10 50
consider reversible reactions, products are not present in the
4 58 4 2.5 25
5 58 4 5 25 beginning of the reaction and either surface reaction, adsorption
6 58 4 15 25 ordesorption as the rate-limiting step. LHHW models consider
7 48 4 10 25 that both reactants are adsorbed while ER models consider that
8 68 4 10 25 only one reactant is adsorbed, while the other is in the liquid
9 78 4 10 25
phase [41]. The derivation of equations for each model
112 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119

Table 2 where –rA is the lauric acid reaction rate and it is equal to
Langmuir–Hinshelwood–Hougen–Watson (LHHW) kinetic models proposed for
expressions given by Tables 2 and 3 for rate equation (r), W is the
esterification reaction of lauric acid with ethanol.
weight of catalyst, Vrxn is the reaction volume, CA0 is the initial
Model Rate-limiting step Rate equation (r) concentration of lauric acid, NA0 are the initial moles of lauric acid,
Definition of k
t is reaction time, Ci is the molar concentration of specie i (Eq. (3)),
h i
LH1 Acid lauric adsorption k ¼ kA Ct k CA K
CC CD
ki are the reaction constants and Ki are the equilibrium constants.
h S KA KB KC KD CB
i
1þK
CC CD C C
þKB CB þKC þKD
The Arrhenius’ equation (Eq. (5)) was used to calculated the
LH2 S KB KC KD CB C D
h i kinetic constants (k). Temperature dependence of Ki is known from
Ethanol adsorption k ¼ kB Ct k CB K
CC CD

h S KA KB KC KD CA
i the chemical reaction isotherm equation (Van’t Hoff equation) and
the definition of Gibbs free energy (DG
), which is given in Eq. (9)
CC CD C C
1þKA CA þK þ Cþ D
LH3 S KA KC KD CA KC KD
h i
Surface reaction k ¼ kS C2t k KA KB CA CB K
CC CD
S KC KD
[43,50,51]. It is assumed that the reaction enthalpy (DH
) and
h i2
LH4 1þKA CA þKB CB þKC þKD
C C reaction entropy (DS
) are both constants within the experimental
C D
h i temperature range [43].
Ethyl laurate desorption k ¼ kC Ct k
KS KA KB KD CA CB CC
K
h
CD C
i   
  
1þKA CA þKB CB þ
KS KA KB KD CA CB CD
þK k DG
DS DH

LH5
h
CD D
i K i ¼ i ¼ exp ¼ exp exp ð9Þ
Water desorption k ¼ kD Ct k
KS KA KB KC CA CB CD
K
ki RT R RT
h i
CC D

1þKA CA þKB CB þKC þ

C
C
KS KA KB KC CA CB
CC Where ki and ki is the constant of direct and reverse reaction,
respectively. As (DS
) is considered constant, the Eq. (9) is
simplified to Eq. (10), which it is similar form to Arrhenius law. This
is the mathematical expression used in this work for the
considering different rate-limiting steps for determination of equilibrium constants of adsorption, surface reaction and desorp-
esterification reaction rate constants is given in Supplementa- tion.
ry information, Appendix A is for LHHW models and  
DH

Appendix B is for ER models. K i ¼ Ai exp ð10Þ
RT
The reaction rate is related with lauric acid conversion (X)
according to Eq. (8) [42]: The objective function (OF) presented in Eq. (7) was minimized
for finding the optimal parameters for each one of the heteroge-
V rxn C A0 dX NA0 dX dX W neous models of Tables 2 and 3; it was done using the fmincon
r ¼ rA ¼ ¼ ! ¼ r ð8Þ
W dt W dt dt NA0
subroutine from Matlab1 . For each model, the root mean square

Table 3
Eley–Rideal (ER) kinetic models proposed for esterification reaction of lauric acid with ethanol.

Model Adsorbed reactant Desorbed product Rate-limiting step Rate equation (r)
Definition of k
h i
ER1 Lauric acid Ethyl laurate Adsorption k ¼ kA Ct k CA K
CC CD

h S KA KC CB
i
CC CD C
1þK þ C
ER2 S KC CB KC
h i
Surface reaction k ¼ kS Ct k KA CA CB  KC KD
C C

h S C
i
C
1þKA CA þKC
ER3 C
h i
Desorption k ¼ kC Ct k
KS KA CA CB CC
K
h i
CD C

KS KA CA CB
1þKA CA þ
ER4 CD
h i
Lauric acid Water Adsorption k ¼ kA Ct k CA K
CC CD

h S KA KD CB
i
C C C
1þK CK DC þKD
ER5 S D B D
h i
Surface reaction k ¼ kS Ct k KA CA CB KC KD
C C

h S D
i
C
1þKA CA þKD
ER6 D
h i
Desorption k ¼ kD Ct k
KS KA CA CB CD
K
h i
CC D

KS KA CA CB
1þKA CA þ
ER7 CC
h i
Ethanol Ethyl laurate Adsorption k ¼ kB Ct k CB K
CC CD

h S KB KC CA
i
C C C
1þK CK DC þKC
ER8 S C A C
h i
Surface reaction k ¼ kS Ct k KB CA CB  KC KD
C C

h S C
i
C
1þKB CB þKC
ER9 C
h i
Desorption k ¼ kC Ct k
KS KB CA CB CC
K
h i
CD C

KS KB CA CB
1þKB CB þ
ER10 CD
h i
Ethanol Water Adsorption k ¼ kB Ct k CB K
CC CD

h S KB KD C A
i
CC CD C
1þK þ D
ER11 S KD CA KD
h i
Surface reaction k ¼ kS Ct k KB CA CB KC KD
C C

h S D
i
C
1þKB CB þKD
ER12 D
h i
Desorption k ¼ kD Ct k
KS KB CA CB CD
K
h i
CC D

KS KB CA CB
1þKB CB þ CC
 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119 113

error (RMSE), normalized root mean square error (NRMSE) and

determination coefficient (R2) were calculated by Eqs. (11)–(13),
respectively [44].
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uP
u Nt
Exp 2
u Xi  X i Mod
t i
RMSE ¼ ð11Þ
Nt

RMSE
NRMSE ¼ ð12Þ
maxðxi Þ  minðxi Þ

Nt

P 2
X i Exp  X i Mod
i
R2 ¼ 1  Nt
 ð13Þ
P Exp 2
X i Exp  X
i

Where Nt is the total data points used for calculation, maxðxi Þ and
minðxi Þ is the largest and smallest value, respectively, for lauric Fig. 3. XRD patterns of PMoV, SiO2@20PMoV, and SiO2.
Exp
acid conversion in experimental data set, and  X is the mean of
experimental data.

Results and discussion

micrographs for SiO2 and SiO2@20PMoV, without porosity in the
500 magnification (Fig. 4). The similitude in the FT-IR spectra, XRD
Catalyts characterization
patterns and SEM images among of include catalysts and silica,
indicates that the PMoV is well disperse into the silica framework.
FT-IR spectra of the synthesized catalysts are similar to the
The nitrogen adsorption/desorption isotherms of the included
pure silica. They present signals which are associated to
catalysts are typical Langmuir isotherms [56], which indicate that
amorphous silica: 3500 cm1 (dSi–OH), 1639 cm1 (dO–H),
the synthezised solid are microporous materials (Figs. S1–S5,
1080 cm1 (nO–Si–O), 951 cm1 (nSi–OH), 797 cm1, and
Supplementary information). Specific surface area (SBET), pore
453 cm1 (dO–Si–O) [49] which mask the characteristic peaks
volume and pore size distribution of silica and catalysts are
of Keggin structure of PMoV (1060 cm1 (nP–Oa), 960 cm1
presented in Table 4. It can be seen that the values of included
(nMo–Od), 864 cm1 (nMo–Ob–Mo), 777 cm1 (nMo–Oc–Mo))
catalysts are similar to pure silica. Lower values are obtained for
[52]. Fig. 2 shows an interval of the spectra of PMoV, SiO2, and the
SiO2@5PMoV and SiO2@10PMoV, which can be attributed to pore
representative catalyst SiO2@20PMoV.
blocking by the remaining PMoV inside silica framework. On the
Fig. 3 shows the X-ray diffraction patterns of PMoV, SiO2, and
other hand, with an increase of PMoV load, the strong compression
SiO2@20PMoV. The bulk PMoV pattern presents the distinctive
during sol–gel process is prevent by the PMoV excess, resulting in
signals of Keggin structure in the intervals 2u = 610o , 1623o , and
an increment in the superficial area, pore volume and size.
2536o [53], although these peaks are not present in the included
The curves acquired by potentiometric titration with n-butyl-
catalysts since are masked by the large diffraction peak of the
amine, allow obtain the maximum acid strength of the surface sites
amorphous silica patterns. SiO2@20PMoV presents the same
and the total number of acid sites. The former is indicated by the
driffraction shape of an amorphous material [54].
initial potential (E0) and the total number of acid sites is given by
The amorphous nature of the silica [55], with the typical laminar
the plateau. The acid strength of catalysts are listed in Table 4. The
morphology and size irregularity are observeed in the SEM
incorporation of PMoV into the silica framework produce a E0
decreament, because the protons can interact with SiO2. The acid
strength of included catalysts increases with the PMoV load
increment. Pure silica and SiO2@5PMoV has strong sites
(100 > E0 > 0 mV) meanwhile the other catalysts have very strong
acid sites (E0 > 100 mV), according to Cid and Pecchi clasification
[57]. This characteristic allow the use of this materials as acid
catalysts in the esterification of lauric acid. Fig. 5 shows the PMoV,
SiO2@20PMoV, and SiO2 titration curves.

Catalytic esterification reaction

Effect of heteropoly acid amount

Fig. 2. FT-IR spectra of PMoV, SiO2@20PMoV, and SiO2.
The effect of the HPA amount in the catalyst has been studied in
the esterification reaction of lauric acid with ethanol at 58
C, using
4 mmol of fatty acid, 10 mL of alcohol and 25 mg of catalyst. Fig. 6
shows the conversion profile of lauric acid for different PMoV
loadings and the respective blank, that is, without catalyst. It is
observed that the fatty acid conversion at specific time increases as
the HPA loading increases from 0 to 15%; however, using more than
15% does not significantly affect the conversion.
114 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119

Fig. 4. SEM 500 micrographs of SiO2@20PMoV (left) and SiO2 (right).

Table 4
Textural porperties and acidity of catalysts.

Sample SBET (m2/g) Pore volume Pore size (Å) E0 (mV)

(cm3/g)
SiO2 440 0.20 18.6 33.2
SiO2@5PMoV 370 0.17 18.8 80.0
SiO2@10PMoV 290 0.14 18.7 279.3
SiO2@15PMoV 495 0.24 19.4 413.1
SiO2@20PMoV 545 0.27 20.0 433.3
PMoV – – – 831.0

Fig. 6. Effect of HPA loading in conversion profile of lauric acid in esterification

reaction. Reaction conditions: 4 mmol lauric acid, 10 mL ethanol, 25 mg catalyst,
58
C.

Fig. 5. Potentiometric titration with n-butylamine of PMoV, SiO2@20PMoV, and SiO2.

Catalyst reuse
The reusability of heterogeneous catalyst is one of its main
advantages compared to homogeneous catalysts. This characteristic
has been studied in the esterification reaction of lauric acid with ethanol
at 58
C by running consecutive reaction cycles under the same reaction
conditions, using 4 mmol of fatty acid, 10 mL of alcohol and 25 mg of
catalyst with a heteropoly acid loading of 20% (SiO2@20PMoV). Fig. 7
shows that the lauric acid conversion during five cycles is very similar,
indicating the high stability of the heterogeneous catalyst.

Effect of chain length on esterification reaction

Fig. 7. Catalytic activity in the esterification reaction of lauric acid with ethanol, in
The effect of the fatty acid has been studied in the esterification five consecutive cycles. Reaction conditions: 4 mmol lauric acid, 10 mL ethanol,
reaction with ethanol at 78
C by 5 h, using 4 mmol of fatty acid,10 mL 25 mg SiO2@20PMoV, 58
C.
 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119
of alcohol and 25 mg of SiO2@20PMoV. Fig. 8 shows the effect of
different fatty acids. Under the same reaction conditions, the
conversion of fatty acid decreases as the length of the fatty acid chain
increases, obtaining the highest conversion (98%) with lauric acid
and the lowest conversion with stearic (18:0) and oleic (18:1) acids,
whose values are 83% and 80%, respectively. This behavior has been
evidenced previously for the esterification reaction of oleic and
acetic acid with ethanol, using Amberlyst 15 as a catalyst [44].
Effect of alcohol
The effect of the alcohol nature has been studied in the
esterification reaction with lauric acid at 78
C by 5 h, using 4 mmol
of fatty acid, 10 mL of alcohol and 25 mg SiO2@20PMoV. Fig. 9 shows
that lauric acid conversion is strongly dependent to the alcohol
nature. It is observed that higher conversion is achieved, in general,
for the longer linear chain alcohols such as isopentanol, n-butanol, n-
propanol and ethanol. By comparison the phenol and the benzyl
alcohol, it is concluded that the location of the hydroxyl group has a
very important effect, because the lauric acid conversion is less
when phenol is used probably due to the aromatic nature of carbon.
In addition, a significant steric effect on the tertiary carbon has been
reported in the literature for the esterification of stearic acid with 1-
and 2-butanol at 150
C, using a montmorillonite-based clay
catalyst [58]. The above can be checked with the results obtained
for butanol isomers, since the lauric acid conversion is 98%, 90% and
44% for n-butanol, s-butanol and ter-butanol, respectively,
Fig. 8. Effect of fatty acid in conversion at 5 h in esterification reaction. Reaction
conditions: 4 mmol fatty acid, 10 mL ethanol, 25 mg SiO2@20PMoV, 78
C.
Fig. 9. Effect of alcohol nature in conversion of lauric acid at 5 h in esterification
reaction. Reaction conditions: 4 mmol lauric acid, 10 mL alcohol, 25 mg SiO2@20P-
MoV, 78
C.
115
corresponding to a primary, secondary, and tertiary alcohol,
respectively. Similar conclusions are obtained for i-propanol and
n-propanol. An important difference is observed between the
conversion with methanol and ethanol, which can be justified by
their boiling point, whose values are approximately 64.7
C and
78.4
C, respectively, which implies that methanol evaporates at the
reaction temperature and there is not enough alcohol in the reaction
mixture for the reaction to be carried out properly.
Effect of catalyst amount
The effect of the catalyst amount, with SiO2@20PMoV, has been
studied in the esterification reaction of lauric acid with ethanol at
58
C, using 4 mmol of fatty acid and 10 mL of alcohol. Fig. 10 shows
that the lauric acid conversion at specific time increases as the
catalyst amount increases from 12.5 to 50 mg, indicating the great
importance of catalyst surface availability for esters formation.
Effect of ethanol amount
The effect of the ethanol amount has been studied in the
esterification reaction of lauric acid with ethanol at 58
C, using
Fig. 10. Experimental (symbols), LH3 kinetic model (continuous lines) and ER3
kinetic model (dashed lines) for conversion profile of lauric acid at different catalyst
amounts. Reaction conditions: 4 mmol lauric acid, 10 mL alcohol, SiO2@20PMoV as
catalyst, 58
C.
Fig. 11. Experimental (symbols), LH3 kinetic model (continuous lines) and ER3
kinetic model (dashed lines) for conversion profile of lauric acid at different ethanol
amounts. Reaction conditions: 4 mmol lauric acid, 25 mg SiO2@20PMoV, 58
C.
116 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119
Fig. 12. Experimental (symbols), LH3 kinetic model (continuous lines) and ER3
kinetic model (dashed lines) for conversion profile of lauric acid at different
temperatures. Reaction conditions: 4 mmol lauric acid, 10 mL alcohol, 25 mg
SiO2@20PMoV.
4 mmol of fatty acid and 25 mg of catalyst (SiO2@20PMoV). Fig. 11
shows that the lauric acid conversion is not significantly different,
at a specific time, for the four ethanol to lauric acid molar ratios
studied. This probably occurs because the molar ratios correspond
to values between 10.7 and 64.2, which are well above the values
reported in literature [44] as 1, 3, and 5, using acetic and oleic acid
with ethanol, which indicates that there is a large excess of alcohol
in the reaction. For other reaction systems, the esterification
increased as the ethanol to fatty acid ratio increased, indicating
that ethanol facilitates fatty acid conversion on the surface of the
catalyst during the reaction [44].
Effect of temperature
The effect of temperature has been studied in the esterification
reaction of lauric acid with ethanol, using 4 mmol of fatty acid,
10 mL of alcohol and 25 mg of catalyst (SiO2@20PMoV). Fig. 12
shows that the lauric acid conversion, at a specific time, increases
when the reaction temperature is increased from 48 to 78
C.
Conversions around 100% are reached at 24 h with all temper-
atures, except for 48
C, which reaches about 60%.
Kinetic modeling
Pseudo-homogeneous model
The optimal values for kinetic parameters of pseudo-homogeneous
model described by Eq. (6) are reported in Table 5 with the
corresponding determination coefficient (R2) of the model, which
corresponds to 85.79%. Some heterogeneous kinetic models are studied
below, which is coherent due to the catalyst used in this work. Figs. S6–
S8 (Supplementary information) show graphically the behavior of the
pseugo-homogeneous model with respect to the experimental data.
The activation energies calculated by pseudo-homogeneous
model, for this esterification reaction catalyzed by different catalysts
are shown in Table 6. The catalyst studied in this work showed Ea for
direct reaction lower than ZnL2 and [Hnmp]HSO4 but higher values
than the other catalysts for both direct and reverse reaction. However,
the results are biased by the not so good fit of this model (R2 = 85.79%).
More reliable results are shown below for heterogeneous models.
Heterogeneous models
Table 7 present the results of LHHW and ER kinetic model fitting
analysis. These results were obtained assuming that equilibrium
constants are constant in the temperature range studied (48–78
C),
Table 5
Kinetic parameters for the pseudo-homogeneous model of
esterification reaction of lauric acid with ethanol.
Parameter Value
A1 (L3.71/g h mol2.71) 124,000
A2 (L2.96/g h mol1.96) 113,000
E1 (kJ mol1) 60
E2 (kJ mol1) 110
α 1.04
β 2.67
g 1.48
d 1.48
R2 (%) 85.79
Table 6
Activation energies for the esterification reaction of lauric acid and ethanol
catalyzed by different catalysts with pseudo-homogeneous model.
Catalyst Ea, direct (kJ mol1) Ea, reverse (kJ mol1) Reference
SiO2@20PMoV 60.00 110.00 This study
ZnL2 67.96 64.31 [36]
[(n-bu-SO3H)MIM] 51.40 [37]
[HSO4]
CTAB-DES 48.78 40.66 [38]
[Hnmp]HSO4a 68.45 [39]
a
Methanol instead of ethanol.
which can be a good estimation and approximation (based on Ref.
[43]) for finding the kinetic parameters through non-linear
regression, because there is a large parameters set. The best LHHW
model to describe the esterification reaction of lauric acid with
ethanol with the heterogeneous catalyst was, slightly better, LH3,
followed by LH2, because LH3 shows the best statistical parameters,
highest R2 (94.7%) and lowest RMSE (0.065) and NRMSE (0.065). This
means that using LHHW kinetic model, the limiting step of the
esterification is the surface reaction as has been reported in the
literature for oleic acid esterification with ethanol catalyzed by
Amberlyst 15 as catalyst [44] or acetic acid esterification with
isopropyl alcohol catalyzed by ion exchange resin [43].
Based on the statistical parameters as mentioned above, the best
ER models to describe the esterification reaction of lauric acid were
ER3 and ER6, which corresponds to desorption of ethyl laurate or
water as rate-limiting step, respectively, when the adsorbed reactant
is lauric acid. These models are so similar due to rate equation
expressions are similar because it was assumed that there were
initially none of the products. In conclusion, model fitting for lauric
acid esterification indicates that the ER model describes the reaction
slightly better than LHHW model, when the equilibrium constants
are considered constant in temperature range.
The two previous models (LH3 and ER3) were taken, one for each
type of kinetic model, to determine the kinetic parameters
considering the variation of the equilibrium constants with
temperature (Eq. (10)). The parameters are shown in Table 8, where
the fitting for the LH3 model (R2 = 96.7%) is improved compared to
corresponding shown in Table 7 considering Ki constants
(R2 = 94.7%). However, for the ER3 model, the improvement is very
slight since R2 is 96.7% instead of 96.0%. Due to the adsorption is an
exothermic process and desorption an endothermic process [59], the
enthalpy changes are negative and positive, respectively, as shown in
Table 8. The energy activation values do not change significantly
between the LH3 and ER3 models in Tables 7 and 8.
The lauric acid esterification with ethanol can be described by either
the LH3 or ER3 models, taking into account the dependence of
equilibrium constants with the temperature. Both models showa higher
fitting to experimental data in comparison with pseudo-homogeneous
model. Figs.10–12 showgraphically the behavior of LH3 and ER3 models
with respect to the experimental data.
 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119 117

Table 7
Kinetic parameters for the LHHW and ER kinetic models of esterification reaction of lauric acid with ethanol.

Model Parameter

A (L g1 h1) Ea (kJ mol1) KA (L mol1) KB (L mol1) KC (mol L1) KD (mol L1) KS R2 RMSE NRMSE
LH1 7.779E+04 37.966 1.270E+04 1.148E-04 4.095E+00 4.095E+00 4.846E+00 0.788 0.127 0.129
LH2 5.003E+05 54.677 1.522E-06 6.040E+03 3.022E+02 9.472E+01 2.165E+01 0.943 0.066 0.067
LH3a 6.123E+07 56.337 2.988E+05 7.490E+03 1.241E+03 1.150E-05 5.211E+05 0.947 0.065 0.065
LH4a 2.748E+05 45.376 1.454E+01 3.260E+06 7.042E+07 1.789E-05 1.180E+04 0.914 0.081 0.082
LH5a 2.748E+05 45.376 1.454E+01 3.260E+06 1.789E-05 7.042E+07 1.180E+04 0.914 0.081 0.082
ER1 2.361E+10 73.608 3.720E+07 – 9.170E+07 – 5.559E+07 0.757 0.136 0.137
ER2 6.949E+05 55.357 4.320E+01 – 1.156E-01 – 1.567E-01 0.949 0.062 0.063
ER3a 3.378E+08 64.954 7.014E-01 – 2.442E+08 – 2.729E-01 0.960 0.055 0.056
ER4 2.361E+10 73.608 3.720E+07 – – 9.170E+07 5.559E+07 0.757 0.136 0.137
ER5 6.949E+05 55.357 4.320E+01 – – 1.156E-01 1.567E-01 0.949 0.062 0.063
ER6a 3.378E+08 64.954 7.014E-01 – – 2.442E+08 2.729E-01 0.960 0.055 0.056
ER7 1.000E+06 56.636 – 1.008E+07 3.586E+05 – 2.792E-06 0.947 0.064 0.064
ER8 2.427E+07 52.239 – 1.301E-01 8.824E-02 – 7.259E+03 0.824 0.116 0.117
ER9a 4.334E+07 59.343 – 5.779E+07 6.475E+07 – 2.858E+00 0.950 0.062 0.063
ER10 1.000E+06 56.636 – 1.008E+07 – 3.586E+05 2.792E-06 0.947 0.064 0.064
ER11 2.427E+07 52.239 – 1.301E-01 – 8.824E-02 7.259E+03 0.824 0.116 0.117
ER12a 4.334E+07 59.343 – 5.779E+07 – 6.475E+07 2.858E+00 0.950 0.062 0.063
a
Units of parameter A: mol g1 h1.

Table 8
Kinetic parameters for the LH3 and ER3 models of esterification reaction of lauric  kS
acid with ethanol.
$
A:S þ B:S ð14cÞ
Parameter LH3 ER3  kS
C:S þ D:S surface reaction
A (mol g1 h1) 6.123E+07 3.378E+08
AA (L/mol) 2.997E+04 1.349E+00
AB (L/mol) 3.523E-06 –
AC (mol/L) 5.183E+03 1.828E+04  kC
AD (mol/L) 3.450E+04 – $
AS 5.176E+03 9.137E+02 C:S ð14dÞ
 kC
Ea (kJ mol1) 56.072 64.824
C þ S desorption of ethyl laurate
DH
A (kJ mol1) 5.996 0.019
DH
B (kJ mol1) 59.354 –
DH
C (kJ mol1) 31.750 17.150
DH
D (kJ mol1) 61.332 –  kD
DH
S (kJ mol1) 15.071 24.531 $
R2 0.967 0.967 D:S ð14eÞ
 kD
RMSE 0.051 0.050
NRMSE 0.051 0.050
D þ S desorption of water

The active sites correspond to Brönsted acid sites associated  ER3 (ethyl laurate desorption as the rate limiting-step)
with available hydrogen on the Keggin structure. These structures
are within the pores or on the surface of silica support through its
interaction with silanols. The presence of Brönsted acid sites have
been demonstrated by pyridine-FTIR in the commercial Keggin  kA
HPAs with similar structure [60]. The LHHW (LH3) and the Eley– $
AþS ð15aÞ
Rideal (ER3) mechanism-based models were proposed as potential  kA
A:S adsorption of lauric acid
reaction mechanisms for esterification of lauric acid with ethanol
over Keggin heteropolyacid loaded in silica framework. The
following reaction steps were proposed, where S represents the
 kS
surface site on catalyst.
$
A:S þ B ð15bÞ
 kS
 LH3 (surface reaction as the rate limiting-step) C:S þ D surface reaction 

 kA
$  kC
AþS ð14aÞ
 kA $
C:S ð15cÞ
A:S adsorption of lauric acid  kC
C þ S desorption of ethyl laurate
The above mechanism-based models are in accordance with
 kB reported for other esterification reactions such as the lauric
$
BþS ð14bÞ acid esterification with ethanol over acid activated montmo-
 kB
rillonite [46], the oleic acid esterification with methanol over
B:S adsorption of ethanol
Amberlyst 46 [42], the esterification reaction of acetic and
118 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119
oleic acids with ethanol over Amberlyst 15 [44] and the lauric
acid esterification with glycerol over 12-tungstophosphoric
acid/SBA-15 [45].
Conclusions
Kinetic models were developed to describe the esterification
reaction of lauric acid and ethanol on heterogeneous catalysts
vanadium Keggin HPA included on a silica framework. This reaction
has special interes in the biodiesel production. Four catalysts with
different PMoV loadings were synthesized by sol–gel procedure,
charactherized and tested as solid catalysts in the esterification of
lauric acid. Best catalytic performance was achieved with SiO2@20P-
MoV, and its reusability was successfully proven in five consecutive
runs. Moreover, the esterification was extended to others fatty acids
and alcohols. Kinetic data with SiO2@20PMoV catalyst were recolected
at different temperatures, acid:alcohol molar ratios, and catalyst
amounts, and used in a pseudo-homogeneous reversible model and
LHHW and ER heterogeneous models. LH3 was the best LHHW model,
with the surface reaction as the rate-limiting step of the lauric acid
esterification, while ER3 was the best ER model, which corresponds to
desorption of ethyl laurate as the rate-limiting step when the adsorbed
reactant is lauric acid. The LH3 and ER3 heterogeneous models showed
a higher fitting to experimental data than pseudo-homogeneous
model (R2 = 96.7% for both models greather than 85.8%), whose
activation energies were 56.1 kJ mol1 and 64.8 kJ mol1, pre-expo-
nential factors were 6.1E + 07 and 3.4E + 08 mol g1 h1, and the
reaction rate constants at 78
C were 0.2791 mol g1 h1 and
0.0768 mol g1 h-1 for LH3 and ER3, respectively.
Declaration of interests
All the authors declare that they have no conflict of interest.
Declaration of Competing Interest
The authors report no declarations of interest.
Acknowledgments
The authors thank to G. Valle, J. Tara, and M. Theiller, for the
experimental measurements. We also thank to UNLP,CONICET, and
CIC for the financial support. VP, PGV, and GPR are members of
CONICET.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.jiec.2020.08.030.
References
[1] X. Tang, S. Niu, S. Zhao, X. Zhang, X. Li, H. Yu, C. Lu, K. Han, J. Ind. Eng. Chem. 77
(2019) 432, doi:http://dx.doi.org/10.1016/j.jiec.2019.05.008.
[2] G.D. Yadav, A.R. Yadav, Chem. Eng. J. 243 (2014) 556, doi:http://dx.doi.org/
10.1016/j.cej.2014.01.013.
[3] X.-X. Han, K.-K. Chen, W. Yan, C.-T. Hung, L.-L. Liu, P.-H. Wu, K.-C. Lin, S.-B. Liu,
Fuel 165 (2016) 115, doi:http://dx.doi.org/10.1016/j.fuel.2015.10.027.
[4] X. Tang, S. Niu, J. Ind. Eng. Chem. 69 (2019) 187, doi:http://dx.doi.org/10.1016/j.
jiec.2018.09.016.
[5] S.Y. Chua, L.A. Periasamy, C.M.H. Goh, Y.H. Tan, N.M. Mubarak, J. Kansedo, M.
Khalid, R. Walvekar, E.C. Abdullah, J. Ind. Eng. Chem. 81 (2020) 41, doi:http://
dx.doi.org/10.1016/j.jiec.2019.09.022.
[6] M. Banchero, G. Gozzelino, Energies 11 (7) (2018) 1843, doi:http://dx.doi.org/
10.3390/en11071843.
[7] F.D. Gunstone, J.L. Harwood, A.J. Dijkstra, The Lipid Handbook, 3rd ed., CRC
Press, Boca Raton, London, 2007.
[8] D.C. Taylor, M.A. Smith, P. Fobert, E. Mietkiewska, R.J. Weselake, 2nd ed.,
Comprehensive Biotechnology, vol. 4, Elsevier, 2011 p. 67.
[9] S. Ganesan, S. Nadarajah, M. Khairuddean, G.B. Teh, Renew. Energy 140 (2019)
9, doi:http://dx.doi.org/10.1016/j.renene.2019.03.031.
[10] A. Gaurav, S. Dumas, C.T.Q. Mai, F.T.T. Ng, Green Energy Environ. 4 (3) (2019)
328, doi:http://dx.doi.org/10.1016/j.gee.2019.03.004.
[11] S. Furuta, H. Matsuhashi, K. Arata, Catal. Commun. 5 (12) (2004) 721, doi:
http://dx.doi.org/10.1016/j.catcom.2004.09.001.
[12] J. Boro, A.J. Thakur, D. Deka, Fuel Process. Technol. 92 (10) (2011) 2061, doi:
http://dx.doi.org/10.1016/j.fuproc.2011.06.008.
[13] N.A. Abas, R. Yusoff, M.K. Aroua, H. Abdul Aziz, Z. Idris, Chem. Eng. J. 382 (2020)
122975, doi:http://dx.doi.org/10.1016/j.cej.2019.122975.
[14] A.I. Tropecêlo, M.H. Casimiro, I.M. Fonseca, A.M. Ramos, J. Vital, J.E.
Castanheiro, Appl. Catal. A Gen. 390 (1–2) (2010) 183, doi:http://dx.doi.org/
10.1016/j.apcata.2010.10.007.
[15] R. Gomes Prado, M.L. Bianchi, E. Gaspar da Mota, S. Silva Brum, J. Henrique
Lopes, M.J. da Silva, Waste Biomass Valorization 9 (4) (2018) 669, doi:http://dx.
doi.org/10.1007/s12649-017-0012-0.
[16] M.A. Hanif, S. Nisar, U. Rashid, Catal. Rev. 59 (2) (2017) 165, doi:http://dx.doi.
org/10.1080/01614940.2017.1321452.
[17] K. Vasi c, G. Hojnik Podrepšek, Ž. Knez, M. Leitgeb, Catalysts 10 (2) (2020) 237,
doi:http://dx.doi.org/10.3390/catal10020237.
[18] Y. Jeon, W.S. Chi, J. Hwang, D.H. Kim, J.H. Kim, Y.-G. Shul, Appl. Catal. B Environ.
242 (2019) 51, doi:http://dx.doi.org/10.1016/j.apcatb.2018.09.071.
[19] L.M. Sanchez, H.J. Thomas, M.J. Climent, G.P. Romanelli, S. Iborra, Catal. Rev.
Sci. Eng. 58 (4) (2016) 497, doi:http://dx.doi.org/10.1080/
01614940.2016.1248721.
[20] O.M. Portilla-Zuñiga, Á.G. Sathicq, J.J. Martínez, S.A. Fernandes, T.R.M. Rezende,
G.P. Romanelli, Sustain. Chem. Pharm. 10 (257) (2018) 50, doi:http://dx.doi.
org/10.1016/j.scp.2018.09.002.
[21] M. Palacio, P.I. Villabrille, V. Palermo, G.P. Romanelli, J. Sol-Gel Sci. Technol. 95
(2020) 321, doi:http://dx.doi.org/10.1007/s10971-020-05239-6.
[22] D.M. Morales, R.A. Frenzel, G.P. Romanelli, L.R. Pizzio, Mol. Catal. 481 (2020)
110210, doi:http://dx.doi.org/10.1016/j.mcat.2018.10.005.
[23] H. Rastegari, H.S. Ghaziaskar, J. Ind. Eng. Chem. 21 (2015) 856, doi:http://dx.
doi.org/10.1016/j.jiec.2014.04.023.
[24] O.D.S. Lacerda, R.M. Cavalcanti, T.M. De Matos, R.S. Angélica, G.N. Da Rocha
Filho, I.D.C. Lopes Barros, Fuel 108 (2013) 604, doi:http://dx.doi.org/10.1016/j.
fuel.2013.01.008.
[25] V. Şimşek, K. Mürtezaog lu, Bilecik Şeyh Edebali Üniversitesi Fen Bilim. Derg. 6
(1) (2019) 91, doi:http://dx.doi.org/10.35193/bseufbd.553967.
[26] R. Sheikh, M. Choi, J. Im, Y. Park, J. Ind. Eng. Chem. 19 (4) (2013) 1413, doi:http://
dx.doi.org/10.1016/j.jiec.2013.01.005.
[27] A. Alsalme, A.A. Alsharif, H. Al-Enizi, M. Khan, S.G. Alshammari, M.A. Alotaibi,
R.A. Khan, M.R.H. Siddiqui, J. Chem. (2018) (2018) 1, doi:http://dx.doi.org/
10.1155/2018/7037461.
[28] B. Hamad, R.O. Lopes de Souza, G. Sapaly, M.G. Carneiro Rocha, P.G. Pries de
Oliveira, W.A. Gonzalez, E. Andrade Sales, N. Essayem, Catal. Commun. 10 (1)
(2008) 92, doi:http://dx.doi.org/10.1016/j.catcom.2008.07.040.
[29] A. de N. de Oliveira, M.A. Barbosa de Lima, L.H. de Oliveira Pires, M. Rosas da
Silva, P.T. Souza da Luz, R.S. Angélica, G.N. da Rocha Filho, C.E.F. da Costa, R.
Luque, L.A. Santos do Nascimento, Materials (Basel). 12 (9) (2019) 1431, doi:
http://dx.doi.org/10.3390/ma12091431.
[30] P. Sudhakar, A. Pandurangan, Mater. Renew. Sustain. Energy 8 (4) (2019) 22,
doi:http://dx.doi.org/10.1007/s40243-019-0160-1.
[31] J. Alcañiz-Monge, B. El Bakkali, G. Trautwein, S. Reinoso, Appl. Catal. B Environ.
224 (2018) 194, doi:http://dx.doi.org/10.1016/j.apcatb.2017.10.066.
[32] J. Alcañiz-Monge, G. Trautwein, J.P. Marco-Lozar, Appl. Catal. A Gen. 468 (2013)
432, doi:http://dx.doi.org/10.1016/j.apcata.2013.09.006.
[33] Q. Zhang, D. Lei, Q. Luo, T. Deng, J. Cheng, Y. Zhang, P. Ma, Period. Polytech.
Chem. Eng. (Ii) (2019) 1, doi:http://dx.doi.org/10.3311/PPch.14788.
[34] Q. Zhang, D. Ling, D. Lei, T. Deng, Y. Zhang, P. Ma, Green Process. Synth. 9 (1)
(2020) 131, doi:http://dx.doi.org/10.1515/gps-2020-0014.
[35] Q. Zhang, D. Ling, D. Lei, J. Wang, X. Liu, Y. Zhang, P. Ma, Front. Chem 8 (March)
(2020) 1, doi:http://dx.doi.org/10.3389/fchem.2020.00129.
[36] E.J.M. De Paiva, V. Graeser, F. Wypych, M.L. Corazza, Fuel 117 (Part A) (2014)
125, doi:http://dx.doi.org/10.1016/j.fuel.2013.09.016.
[37] Z. Wei, F. Li, H. Xing, S. Deng, Q. Ren, Korean J. Chem. Eng. 26 (3) (2009) 666,
doi:http://dx.doi.org/10.1007/s11814-009-0111-0.
[38] K. Li, X. Chen, W. Xue, Z. Zeng, S. Jiang, Int. J. Chem. Kinet. 51 (5) (2019) 329, doi:
http://dx.doi.org/10.1002/kin.21256.
[39] B. Han, W. Zhang, F. Yin, S. Liu, X. Zhao, J. Liu, C. Wang, H. Yang, R. Soc. Open Sci.
5 (9) (2018) 180672, doi:http://dx.doi.org/10.1098/rsos.180672.
[40] Q. Zhang, T. Yang, X. Liu, C. Yue, L. Ao, T. Deng, Y. Zhang, RSC Adv. 9 (29) (2019)
16357, doi:http://dx.doi.org/10.1039/C9RA03209F.
[41] J. De Araújo Gonaçalves, A.L.D. Ramos, L.L.L. Rocha, A.K. Domingos, R.S.
Monteiro, J.S. Peres, N.C. Furtado, C.A. Taft, D.A.G. Aranda, J. Phys. Org. Chem. 24
(1) (2011) 54, doi:http://dx.doi.org/10.1002/poc.1701.
[42] O. Ilgen, Fuel Process. Technol. 124 (2014) 134, doi:http://dx.doi.org/10.1016/j.
fuproc.2014.02.023.
[43] Y. Liu, J. Liu, H. Yan, Z. Zhou, A. Zhou, ACS Omega 4 (2019) 19462, doi:http://dx.
doi.org/10.1021/acsomega.9b02994.
[44] A. Shahid, Y. Jamal, S.J. Khan, J.A. Khan, B. Boulanger, Arab. J. Sci. Eng. 43 (11)
(2018) 5701, doi:http://dx.doi.org/10.1007/s13369-017-2927-y.
[45] P. Hoo, A.Z. Abdullah, Ind. Eng. Chem. Res. 54 (32) (2015) 7852, doi:http://dx.
doi.org/10.1021/acs.iecr.5b02304.
[46] P.R.S. Dos Santos, F. Wypych, F.A.P. Voll, F. Hamerski, M.L. Corazza, Fuel 181 (x)
(2016) 600, doi:http://dx.doi.org/10.1016/j.fuel.2016.05.026.
 L.A. Gallego-Villada et al. / Journal of Industrial and Engineering Chemistry 92 (2020) 109–119
[47] L.A. Petrov, Y. Alhamed, A. Al-Zahrani, M. Daous, Chin. J. Catal. 32 (6–8) (2011)
1085, doi:http://dx.doi.org/10.1016/S1872-2067(10)60225-2.
[48] F. Kern, S. Ruf, G. Emig, Appl. Catal. A Gen. 150 (1) (1997) 143, doi:http://dx.doi.
org/10.1016/S0926-860X 96 00286-4.
[49] V. Palermo, Á. Sathicq, T. Constantieux, J. Rodríguez, P. Vázquez, G. Romanelli, Catal.
Lett. 145 (4) (2015) 1022, doi:http://dx.doi.org/10.1007/s10562-015-1498-3.
[50] Chemistry LibreTexts. The Van’t Hoff Equation. Available at: https://chem.
libretexts.org/. (Accessed 11 January 2020).
[51] L.E. Revell, B.E. Williamson, J. Chem. Educ. 90 (8) (2013) 1024, doi:http://dx.doi.
org/10.1021/ed400086w.
[52] L.R. Pizzio, P.G. Vázquez, C.V. Cáceres, M.N. Blanco, Appl. Catal. A Gen. 256 (1–
2) (2003) 125, doi:http://dx.doi.org/10.1016/S0926-860X(03)00394-6.
[53] A. Popa, V. Sasca, E.E. Kiss, R. Marinkovic-Neducin, M.T. Bokorov, I. Holclajtner-
Antunovi c, Mater. Chem. Phys. 119 (3) (2010) 465, doi:http://dx.doi.org/
10.1016/j.matchemphys.2009.09.026.
119
[54] J.R. Martínez, S. Palomares-Sánchez, G. Ortega-Zarzosa, F. Ruiz, Y. Chumakov,
Mater. Lett. 60 (29–30) (2006) 3526, doi:http://dx.doi.org/10.1016/j.mat-
let.2006.03.044.
[55] E. Papirer, Adsorption on Silica Surfaces, Marcel Dekker, New York, 2000.
[56] J.D. Wright, N.A.J.M. Sommerdijk, Sol–gel Materials Chemistry and Applica-
tions, Taylor & Francis Group, Boca Raton, 2001.
[57] R. Cid, G. Pecchi, Appl. Catal. 14 (1985) 15, doi:http://dx.doi.org/10.1016/S0166-
9834(00)84340-7.
[58] S. Bouguerra Neji, M. Trabelsi, M. Frikha, Energies 2 (4) (2009) 1107, doi:http://
dx.doi.org/10.3390/en20401107.
[59] H. Lynggaard, A. Andreasen, C. Stegelmann, P. Stoltze, Progress in Surface
Science 77 (2004) 71. https://doi.org/10.1016/j.progsurf.2004.09.001.
[60] Y. Yang, G. Lv, W. Guo, L. Zhang, Microporous Mesoporous Mater. 261 (2018)
214, doi:http://dx.doi.org/10.1016/j.micromeso.2017.11.018.